1. Introduction
The existing literature has extensively explored democracy, which primarily focuses on the factors for determining democracy. The determinants of democracy have been identified, and the relationship democracy has with most of its determinants has been clearly established. However, the link between religion and democracy has remained mixed. In particular, Islam receives relatively more attention than other religions, and its relation to democracy has been highly debatable. However, research on whether and how Islam has bearings on the dynamic process of democracy is lacking. This study aims to fill this research gap.
Promoting democracy has been of major global concern, which culminated in the recent Jasmine revolution and the ensuing Arab Spring. Consequently, the Islamic awakening and uprising has rekindled the debate on the relationship between religion and democracy. Considerable attention has been paid to the issue of whether Islam is the reason for the absence of democracy or the low level of democracy in Islamic countries. However, the issue of whether Islamic countries that have recently experienced revolutions would successfully transform into democratic countries, or whether their religious affiliation (i.e., Islam) would hamper their democratic development as widely believed, has yet to be examined. Despite the criticism of Islam as a hindrance to democratic development, Islam should be open to democracy. This idea is bolstered by the fact that countries that have recently undergone a series of democratic movements (e.g., Indonesia, Pakistan, Tunisia, Egypt, Libya, and Morocco) are primarily populated by Muslims, and Muslims have been involved in the democratic movement (e.g., the Muslim brotherhood fighting for democracy and freedom) – otherwise, democratic movements would not ensue. In fact, although a low level of democracy is widely recognized among Islamic countries, there is nevertheless a strong desire for democracy. Such a desire should motivate democratic development in Islamic countries when compared to non-Islamic countries that are currently more democratic (Maseland and van Hoorn, Reference Maseland and van Hoorn2011). However, a lower level of democracy in Islamic countries means that Islamic countries have more room for improvement; therefore, these countries have greater potential to reach a higher level of democracy more rapidly based on the concept of convergence (Verdier, Reference Verdier1998). Hence, the dynamic process of democracy, or democratic adjustment, should be faster in Islamic countries than in non-Islamic countries.
The majority of related studies merely provide literary arguments on the relationship between Islam and democracy, and any empirical work has been retrospective in nature. Thus, this study contributes to the existing literature by exploring the issue of whether and how Islam has any bearing on the dynamic process of democracy. It likewise provides policy recommendations on the manner of promoting democracy more effectively in Islamic countries as opposed to non-Islamic countries. To the best of our knowledge, this research issue remains unexplored in the existing literature.Footnote 1 However, this issue is worth exploring because our results will verify the validity of the view that is generally held by the western world, that is, Islamic countries are unlikely to develop democracy, particularly after the occurrence of Arab Spring when the newly established democracy has been caught in a conundrum. In fact, attributing any ostensibly low level of democracy to Islam alone may be inappropriate. Any pessimism about the future democratic development in Islamic countries can be attributed to other historical and cultural factors that are deeply rooted in the politics of Islamic countries (Kubba et al., Reference Kubba, Muqtedar, Mahmood and Neil2002; Donno and Russett, Reference Donno and Russett2004; Fish, Reference Fish2002; Stepan and Robertson, Reference Stepan and Robertson2003; Weiffen, Reference Weiffen2004). Without solid empirical investigation and clarification, any misconception about the relationship between Islam and democracy is likely to persist, and any existing conflict between Islamic and non-Islamic countries is likely to intensify. This condition prompts and necessitates further clarification, which is the purpose of the current study.
This study uses comprehensive aggregate data from 17 Asian countries (i.e., six Islamic countries and 11 non-Islamic countries) from 1996 to 2010. The results reveal that religion indeed plays a role in determining democracy. More specifically, the level of democracy in Islamic countries is generally lower than that in non-Islamic countries as commonly believed. Nevertheless, democratic adjustment is faster in Islamic countries than in non-Islamic countries. Moreover, the results suggest that Islamic countries can reach a higher level of democracy by improving education, minimizing the gender gap, controlling population growth, or becoming oil exporters similar to non-Islamic countries. However, Islamic countries own a unique advantage in improving democracy, that is, they can effectively promote democracy through urbanization, which turns out to be an ineffective strategy for non-Islamic countries.
The remainder of this paper is organized as follows. A review of the literature that leads to the study hypotheses is initially presented. The research data and methodology are then described, and the empirical results are analyzed. The final section of the paper concludes the study.
2. Literature review
Democracy in Asia
According to Croissant (Reference Croissant2004), the democratic trend in Asia can be classified based on geographic locations. Northeast Asian countries tend to exhibit democratic consolidation, whereas South and Southeast Asian countries are likely to experience democratic stagnation and retreat, respectively. Overall, the majority of Asian countries lag behind western countries in terms of democracy. Although several Asian countries gradually progressed toward democracy during the democratization wave in the 1990s, they seemed to superficially establish an electoral system without realizing the spirit of democracy. The slow or impeded democratic development in these countries can be attributed to the lack of the essential elements of democracy, such as the rule of law, civil society, and accountability (Hood, Reference Hood1998). Consequently, after the incumbent in many countries steps down, the subsequent elections are typically tainted with violence, and military intervention is involved, resulting in the stagnation of democracy. The Philippines and Thailand are good examples of this case (Croissant, Reference Croissant2004).
Any democratic deficit, stagnation, or retreat in Asian countries can also be attributed to the unique Asian cultures, which profoundly influence political development. In particular, Confucianism, considered as a religion by previous research (Taylor, Reference Taylor2014), has the most extensive effect. Huntington (Reference Huntington1991) argued that Confucianism and the significance of democracy are contradictory in the sense that the fundamental ideological element in Confucianism is either no democracy or anti-democracy. Such ideology that is deeply rooted in Asia likely hinders the transition to a truly democratic political regime (Hood, Reference Hood1998). However, similar to other religions, Confucianism may not be entirely detrimental to democratic development because it is characterized by multi-cultural tolerance and is accepting of the idea and values of western culture (Huntington, Reference Huntington1991). Such a characteristic should be at least open to, if not favorable toward, democratic development in Asia (Hu, Reference Hu1997).
The openness of Confucianism to democracy is best exemplified by Japan where Confucianism similarly takes root in its culture. The success of Japan in achieving the highest level of democracy in Asia can be attributed to its active pursuit of westernization and modernization after the Meiji Restoration, which laid the foundations of democratization and helped establish a fully westernized and modernized country (Bowen, Reference Bowen1984). Moreover, after World War II, Japan developed a policy of secularism, which successfully facilitated democratic development (Paul, Reference Paul2005). Hence, although Confucianism influences Japan as in other Asian countries, the willingness of Japan to be westernized is likely the key to its success in realizing full democracy.
In addition to Confucianism, Islam is believed to contribute to ‘undemocracy’ or a low level of democracy in several Muslim-populated Asian countries (Huntington, Reference Huntington1991; Croissant, Reference Croissant2004). However, the relationship between religion and democracy is complicated and mixed based on previous research (Halliday, Reference Halliday1996; Yetiv, Reference Yetiv1997; Barro, Reference Barro1999; Huntington, Reference Huntington1991; Lipset, Reference Lipset1994; Stepan, Reference Stepan2000; Minkenberg, Reference Minkenberg2007; Bloom and Arikan, Reference Bloom and Arikan2012, Reference Bloom and Arikan2013; Fradkin, Reference Fradkin2000; Reichley, Reference Reichley1986). Islam, as one of major religions in the world, should be no exception in this respect. Hence, to ensure fairness and objectivity, the relationship between Islam and democracy should be reconsidered and reexamined.
Islam and democracy
The emergence of Islam and Islamic countries in the seventh century is primarily attributed to the desire of national rulers to unite people through religious ideology, such that people would respect and obey those in power, thus successfully establishing an autocratic empire (Borooah and Paldam, Reference Borooah and Paldam2007). For centuries, these countries adopted and practiced Islamic law (Sharia), whose legal code is derived from the Koran, the Muslim bible. Enforcing Sharia to regulate people is apparently well intentioned in terms of attaining social order. Nevertheless, the legal system practiced in Islamic countries is considered undemocratic in nature, compelling Muslims to live in an undemocratic environment; this condition has caused difficulty in building democracy (Esposito and Mogahed, Reference Esposito and Dalia2008). In other words, Islam is more than a religion because its doctrine provides a set of guidelines for national governance to ensure that governments act according to the commandments (norms) of God; national citizens are expected to abide by these guidelines, and they risk severe punishments in case of non-compliance (Lipset, Reference Lipset1994). The issue that emerges is that several concepts underlying Sharia are considered incompatible with democracy. For instance, men and women are treated unequally based on Sharia, and this practice conflicts with the fundamentals of modern democracy that highlights equality and fairness (The Economist, 2010). Such a patriarchal social system that is inherent and prevalent in Islamic countries has been regarded as the primary contributor to authoritarianism (supremacy) and the low level of democracy (Lipset, Reference Lipset1994).
Even if Sharia is indeed undemocratic from the perspective of western countries, most people in Sharia-practicing countries are devout Muslims, who highly support the implementation of Sharia and never question its legitimacy (Mogahed, Reference Mogahed2006); that is, Sharia law is the only recognized norm of life that regulates all of the aspects of Muslim life. Hence, Muslims who have been extensively exposed to such an environment are unlikely to demand a change by adopting a democratic system (Bukay, Reference Bukay2007). Such complacency with their existing political system can further ensure a lower level of democracy in Islamic countries than in non-Islamic countries.
Furthermore, the experience of being colonized by the western power in the early twentieth century likely contributes to the persistent democratic deficit in Islamic countries. More specifically, Muslims feared that the introduction of the western culture (i.e., modernization and democracy) during the colonial era would threaten their imperialism and culture. Islamic culture has its own religious and political norms; thus, Muslims are reluctant to be influenced by the western culture (Hunter, Reference Hunter2009). To avoid or reduce the influence of western powers on politics, economics, and culture in Islamic countries, Muslims initiated the campaign of reviving the Islamic culture, evoking Islamic nationalism to resist western colonization, democratization, and modernization (Somer, Reference Somer2007). Several Islamic countries remain hostile toward the West because of their past colonial experience. Considering that democracy originated in the West, the hostility of Islamic countries toward the West implies that Muslims are hostile toward democracy, causing difficulty in the development of a democratic system (Farooq, Reference Farooq2011). To resist western colonization, Islamic countries consolidated their military regime and applied any modernization absorbed from the West into military power. This approach resulted in the current militarization of politics in several Islamic countries, which, in turn, reinforced their authoritarian regime and further inhibited the growth of democracy (Stein, Reference Stein2012).
Finally, given the extreme resilience of an authoritarian regime that has been driven by cultural or historical factors for centuries, any democratization effort may be thwarted or any promotion of democracy may be expectedly ineffective or unproductive (Hashemi, Reference Hashemi2003).
In addition to the preceding arguments, previous studies have provided empirical evidence, albeit limited, indicating a negative relationship between Islam and democracy. For instance, the level of democracy has been found to be lower in Islamic countries than in other countries, and Islamic countries encounter difficulty in developing democracy likely because of the high association between Islam and authoritarianism (Barro, Reference Barro1999; Clague et al., Reference Clague, Gleason and Knack2001; Fish, Reference Fish2002). Moreover, after the 9/11 terrorist attacks in the United States, numerous people started to think, or became more convinced, that Islam disfavors democracy and freedom, given that the alleged terrorists are from Islamic countries (Berman, Reference Berman2003; Dalmasso and Cavatorta, Reference Dalmasso and Cavatorta2013). In sum, any negative relationship between Islam and democracy can be attributed to Islam and the factors related to it.
As previously mentioned, in addition to Islam, the legal environment and historical background unique to Islamic countries can also cause difficulty in adopting democratic institutions and realizing the democratic spirit; thus, the level of democracy has remained lower in Islamic countries than in non-Islamic countries. However, similar to other religions, Islam is multi-vocal, whose doctrines can be theoretically used in promoting democracy (Anderson, Reference Anderson2004; Tessler, Reference Tessler2002). Although abundant research has reported the negative relationship between Islam and democracy, considering the potential positive relationship between Islam and democracy, the net relationship between Islam and democracy is ambiguous, depending on the relative magnitude of these two opposing relationships. However, if the negative relationship between the preceding Islam-related factors and democracy is more overwhelming than the positive relationship between Islam and democracy, then the net relationship between Islam and democracy is negative. Based on this analysis, the following hypothesis is proposed:
Hypothesis 1 (H1): The level of democracy is lower in Islamic countries than in non-Islamic countries.
Islam and democratic adjustment
Despite the criticism of Islam as a hindrance to democracy, the relationship between Islam and democracy deserves reconsideration based on several studies. For instance, Mogahed (Reference Mogahed2006) revealed the absence of conflict between Islamic law and democracy based on poll results. In addition, Tessler (Reference Tessler2002) used the Middle East countries as the study sample and reported little relationship between Islam and political orientations, suggesting that Islam neither hinders nor fosters democratic orientations (Ciftci, Reference Ciftci2010). Moreover, any negative relationship between Islamic attachment and support for democracy appears to be driven by Muslim women. Finally, any low level of democracy has been determined to be more related to systemic variables in Islamic countries, such as country, culture, and historical factors (Tessler, Reference Tessler2002; Hashemi, Reference Hashemi2003; Haklai, Reference Haklai2009; Heydemann and Leenders, Reference Heydemann and Leenders2011).Footnote 2 Based on the preceding evidence, any documented negative relationship between Islam and democracy may be attributed to a particular sample coverage or failure to control relevant variables, such as gender and country-specific variables. In other words, the real relationship between Islam and democracy remains undetermined. In fact, any negative relationship between Islam and democracy could be the cumulative consequence of the aforementioned factors rather than resulting from Islam only. Additionally, any historically low level of democracy in Islamic countries does not imply that such negative relationship between Islam and democracy will endure, particularly when we look into the future, noting that the Muslim world has been undergoing a series of democratic movements.
The preceding arguments and findings indicate that any conception on anti-democracy or undemocracy in relation to Islam can be biased and non-substantiated. Although the negative relationship between Islam and democracy has obtained empirical support, the underlying methodology is subject to scrutiny, and the related empirical research is limited. More specifically, in addition to the preceding problems, and missing variables (e.g., gender, colonial experience, culture) that could yield biased results, model specification and research focus should be carefully reconsidered. The controversial relationship between Islam and democracy and any negative relationship may be attributed to Islam and other enduring Islam-related factors; thus the level of democracy, as examined in previous studies, should be superceded by focusing on the speed of democratic adjustment. The underlying reason is that history shows that no authoritarian regime lasts forever, and democracy is expected to spread to every corner of the world (Tocqueville, Reference Tocqueville, Mansfield and Winthrop2000). Moreover, a low level of democracy does not imply slower democratic adjustment or minimal opportunity to achieve a higher level of democracy. For instance, Indonesia and Turkey are typical Islamic countries in Asia, which used to be under authoritarian rule. However, these countries have successfully transformed into democracies, probably because Islam in these countries is less powerful and their governments allow for more interaction between Islam and the political parties rather than putting divine law above secular law (Buehler, Reference Buehler2009). Given that previous studies are essentially backward looking and emphasize static democracy in Islamic countries, the adoption of a forward-looking stance by future researchers is worthwhile when examining the relationship between Islam and democracy. The reason is that democracy should be continual and dynamic rather than sporadic and static (Rustow, Reference Rustow1970). Hence, the current study makes an unprecedented attempt by examining the issue of whether and how the speed of democratic adjustment differs between Islamic and non-Islamic countries to provide new insight into the perennially debated research issue.
As previously mentioned, the low level of democracy in Islamic countries may be related to social, historical, and cultural factors rather than Islam only (Tessler, Reference Tessler2002; Haklai, Reference Haklai2009; Heydemann and Leenders, Reference Heydemann and Leenders2011). Hence, Islam should be further reexamined to ascertain whether it indeed deters democratic development as extensively documented in the existing literature. In fact, a few studies provide evidence indicating that Islam is not inimical but friendly towards and even fosters democracy. For instance, from the perspective of the scripture, Islam includes the concepts of shura, ijma, and ijtihad, which denote consultation, consensus, and independent reasoning, respectively. All of these concepts are the building blocks for democracy, suggesting that democratic development is feasible in Islamic countries (Anderson, Reference Anderson2004). In addition, the recent successful transformation in Islamic countries of several enduring authoritarian regimes into democratic ones, as well as several democratic movements initiated by Islamic organizations, at least signifies that Islam facilitates democratization (Hofmann, Reference Hofmann2004). Otherwise, Muslims would not pursue and realize democracy. In other words, these democratic events disprove the conception that Islam is undemocratic or anti-democracy. Moreover, Islam may not be necessarily less democratic than other religions. For instance, Muslims are found to be more supportive of democracy than non-Muslims, such as those who are affiliated with Eastern Orthodox Christianity (Hofmann, Reference Hofmann2004). Furthermore, given that any democracy established in Islamic countries remains premature and that democracy is an unstoppable trend (Tocqueville, Reference Tocqueville, Mansfield and Winthrop2000), Islamic countries should have considerable room for promoting democracy and should rapidly undertake democratic adjustment as opposed to non-Islamic countries from the perspective of convergence (Verdier, Reference Verdier1998). That is, Islamic countries should have the potential to achieve significant progress in democratic development as opposed to non-Islamic countries that are already highly democratic. Finally, Maseland and van Hoorn (Reference Maseland and van Hoorn2011) revealed a negative relationship between the level of democracy and aspirations for democracy. Given the observed level of democracy being generally lower in Islamic countries than in non-Islamic countries, this finding implies that Muslims have stronger democratic aspirations, which should intensify the propagation of democracy, especially after any incidence that triggers a major democratic reform (Dalmasso and Cavatorta, Reference Dalmasso and Cavatorta2013). Based on the preceding reasoning, the following hypothesis is proposed:
Hypothesis 2 (H2): Democratic adjustment is faster in Islamic countries than in non-Islamic countries.
3. Methodology
Data
The study sample consists of 17 Asian countries (Japan, Taiwan, South Korea, China, India, Singapore, Thailand, Indonesia, Philippines, Russia, Turkey, Malaysia, Vietnam, Pakistan, Sri Lanka, Saudi Arabia, and Kuwait), and the study period spans from 1996 to 2010. This time span is selected because it corresponds to the starting year for the data on democracy (i.e., voice and accountability). Asian countries are selected because of their higher religious diversity compared with western countries where Christianity and Catholicism tend to dominate. In addition, the level of democracy in Asian countries generally lags behind that in western countries (e.g., European and North American countries); such a context makes the examination of democracy in Asian countries worthwhile and interesting, and prompts the research on whether a perceivably low level of democracy is related to religious distribution in Asia. Furthermore, high variation in the level of democracy among countries in Asia, as opposed to the West, similarly allows for richer testing.
In sum, the selection of Asian countries as the study sample is justified by heterogeneity in the level of democracy and religion. In addition, the study sample holds the advantage of not including Islamic countries experiencing an Arab Spring or any similar political upheaval in the sample period. Thus, results based on the study sample should be reliable because such results did not arise from a dramatic shift in the political regime.
The data on democracy (Voice and Accountability (VA)) are derived from the Worldwide Governance Indicators databank.Footnote 3 VA is an index that captures ‘perceptions of the extent to which a country's citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association, and a free media’ (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011).Footnote 4 The values of VA range from –2.5 to 2.5; a high value of VA indicates a high level of democracy in a given country.
To test our hypotheses, we categorize the sampled countries into different religious groups. First, we determine the major religion in a given country using religious demography from the 2012 Report on International Religious Freedom provided by the US Department of State.Footnote 5 Major religion in a given country is defined as a religion believed by the highest percentage (more than 50%) of national citizens. If the percentage of people believing in a religion is less than 50% in a given country, then that country is regarded as having no major religion. Table 1 summarizes the cross-country religious demography and major religions. Six countries (i.e., Indonesia, Kuwait, Malaysia, Pakistan, Saudi Arabia, and Turkey) are classified as Islamic countries because more than 50% of their respective populations believe in Islam. The remaining 11 countries (i.e., China, India, Japan, Philippines, Russia, South Korea, Singapore, Sri Lanka, Thailand, Taiwan, and Vietnam) are non-Islamic countries.
Table 1. Summary of religious demography and major religions by country
Notes: Religious demography is adapted from International Religious Freedom Report for 2012, US Department of State.
* According to Vietnam National Administration of Tourism, ‘over 70% of the population of Vietnam are either Buddhist or strongly influenced by Buddhist practices’, http://www.vietnamtourism.com/e_pages/country/overview.asp.
Non-Islamic countries are further classified into five country groups with different belief types, namely, polytheism, diverse belief, Buddhism, Christianity, and atheism. Given that major religions in India, Japan, and Taiwan (Hinduism, Shinto, and folk religion, respectively) are polytheistic in nature, we classify these three countries into the polytheistic country group. In addition, although no major religions exist in South Korea and Singapore, religious demography shows that populations with religious beliefs account for more than 50% in each country. Thus, we classify both countries as countries with diverse beliefs. In addition, we classify Sri Lanka, Thailand, and Vietnam as Buddhist countries because Buddhism is their major religion. Furthermore, although major religions in the Philippines and Russia are reportedly Catholic and Russian Orthodox, both countries are classified as Christian countries because Catholicism and Russian Orthodox are regarded as branches of Christianity; both religions similarly follow the Bible and Jesus Christ. Finally, China is classified as an atheist country, not only because it is officially atheistic but also because the religious population accounts for substantially less than half (31.4%) of the total population. Table 2 summarizes the finalized classification of countries based on belief types.
Table 2. Classification of countries into Islamic and non-Islamic countries, further divided into five religious groups
Figure 1 presents the secular trend of level of democracy for Islamic and non-Islamic countries. The level of democracy is generally lower in Islamic countries than in non-Islamic countries during the entire sample period. The results of a t-test for the mean values and the median test further indicate that the level of democracy for Islamic countries is significantly lower than that for non-Islamic countries (p-value = 0 for both tests). Results support H1 and concur with conventional wisdom that Islamic countries have a lower level of democracy compared with non-Islamic countries. However, the level of democracy apparently shows an upward trend for Islamic countries, as opposed to non-Islamic countries where a downward trend is observed for the entire sample period. The graphic pattern observed in Figure 1 appears to support H2. However, advanced regression analysis is required to further test H2 and verify whether democratic adjustment is faster in Islamic countries than in non-Islamic countries.
Figure 1 Level of democracy for Islamic and non-Islamic countries, 1996–2010
Table 3 presents the correlation matrix and the variance inflation factors (VIFs) of the variables used in this study. VA is related to other variables, thereby justifying the inclusion of these variables as independent variables in the regression analysis. More specifically, VA is positively correlated with per capita gross domestic product (GDP), primary schooling (SCHOOL), and urbanization (URBAN), whereas VA is negatively correlated with gender gap in primary schooling (GAP), population (POP), and being an oil-exporting country (OIL), thus concurring with previous research findings.Footnote 6 The correlation coefficients for variables other than VA are generally not excessively high, except for the correlation coefficient of ln(GDP) vs. URBAN, and several are statistically insignificant (e.g., SCHOOL vs. GAP, SCHOOL vs. URBAN, SCHOOL vs. OIL, GAP vs. ln(POP), and GAP vs. OIL). In addition, the VIFs are generally low, ranging from 1.07 to 4.12; thus, the concern about multicollinearity can be alleviated.
Table 3. Correlation matrix and variance inflation factors
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. VIF indicates the values of variance inflation factors. * stands for 5% significant.
Table 4 presents the mean and median values of VA and its determinants, as well as the number of observations across countries. The sample consists of 235 country-year observations. Each country yielded 14 observations, except Taiwan and Thailand where the number of observations are 12 and 13, respectively. The mean and median values of VA and its determinants exhibit wide variation. The mean (median) values of VA range from –1.57 (–1.57) (Saudi Arabia) to 0.97 (0.98) (Japan). The mean and median values of VA for the entire sample are –0.30 and –0.28, respectively. The top three countries in terms of GDP per capita are Japan, Kuwait, and Singapore, whereas countries with the lowest GDP per capita are Vietnam, Pakistan, and India. Despite the documented relationship between GDP and democracy, rich countries do not necessarily experience higher levels of democracy than the others. Saudi Arabia, for instance, ranks seventh in terms of GDP per capita in the study sample of 17 countries, despite having the lowest level of democracy. The mean and median values of SCHOOL range from five to six years, except for Kuwait and Russia where the mean and median values of SCHOOL are below five years. GAP is lowest (i.e., 0) in Japan, Taiwan, and the Philippines, whereas it is largest in Sri Lanka where the mean and median values are 0.79 and 1, respectively. URBAN is highest in Singapore where the mean and median values are 1.00, whereas it is the lowest in Sri Lanka where the mean and median values are 0.15. The top three countries in terms of POP are China, India, and Indonesia, whereas the three least populated countries are Kuwait, Singapore, and Sri Lanka. Of the entire sample, only Saudi Arabia, Kuwait, and Indonesia are oil exporters.Footnote 7
Table 4. Descriptive statistics of democracy and its determinants by country, 1996–2010
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. The number of years is denoted by n.
Table 5 presents the descriptive statistics (mean, median, maximum, minimum, standard deviation, and number of observations) of democracy and its determinants for Islamic and non-Islamic countries to facilitate comparison between these two country groups. Islamic and non-Islamic countries account for 35.7% and 64.3% of total observations, respectively. We performed a t-test and median test to examine any difference in the values of variables between the two country groups by comparing non-Islamic countries with Islamic countries. The test statistics for VA are significantly positive, indicating that the level of democracy in non-Islamic countries is significantly higher than that in Islamic countries. The results thus support H1. In addition, the test statistics for URBAN are significantly negative, indicating that the urbanization rate in Islamic countries is significantly higher than that in non-Islamic countries. The test statistics for POP are significantly positive, thus indicating that the population in non-Islamic countries is significantly higher than that in Islamic countries. The test statistics for GDP, SCHOOL, and GAP are insignificant, indicating no significant difference in GDP per capita, primary schooling, and gender gap in primary schooling between Islamic and non-Islamic countries. Finally, the test statistics for OIL are significantly negative, concurring with the observation in Table 4 that all oil exporters are Islamic countries.
Table 5. Descriptive statistics for Islamic and non-Islamic countries, 1996–2010
Notes: This table presents mean, median, maximum (max), minimum (min), and standard deviation (sd) values as well as number of observations (N) for different country groups. Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. * and ** indicate significance at 5% and 1% levels, respectively.
To obtain further insight into the research question, we examined the descriptive statistics (mean, median, maximum, minimum, and standard deviation values as well as number of observations) of democracy and its determinants for Islamic and non-Islamic countries divided into five country groups based on belief types (i.e., polytheism, diverse beliefs, Buddhism, Christianity, and atheism) (Table 6). Islamic countries account for the largest number of observations (84); by contrast, the atheist country group consisting only of China shows the smallest number of observations (14). The t-test results indicate that the value of VA for polytheist nations is significantly larger than that for nations with diverse beliefs (t-value = 5.238, p-value = 0.000). VA is significantly larger for nations with diverse beliefs than for Christian nations (t-value = 5.591, p-value = 0.000), and VA is significantly larger for Christian nations than for Buddhist nations (t-value = 1.976, p-value = 0.026). In addition, VA is higher for Buddhist nations than for Islamic nations but with an insignificant difference (t-value = 0.930, p-value = 0.822). Finally, VA for Islamic nations is significantly larger than for the atheist country (i.e., China) (t-value = 6.183, p-value = 0.000). Thus, religions sorted by level of democracy (from high to low) are as follows: polytheism, diverse beliefs, Christianity, Buddhism, Islam, and atheism. Moreover, the mean value of VA for countries with polytheism, diverse beliefs, and Christianity is higher than the overall mean value of VA (–0.301), whereas the mean value of VA for countries with Buddhism, Islam, and atheism is lower than the overall mean value of VA. Furthermore, excluding atheism associated with only one country in the study sample, polytheist countries show the highest value of VA, whereas Islamic countries exhibit the lowest value of VA. Even if the atheist country is not excluded, Islamic countries have the lowest minimum value of VA (–1.773) among all country groups. The preceding observations lend additional support for H1. As for the determinants of democracy, based on the observation of the corresponding mean and median values, GDP per capita is the highest in countries with diverse beliefs and the lowest in Buddhist countries. The gender gap of primary schooling is the largest in the atheist country and the lowest in Christian countries. The atheist country has the largest population, whereas countries with diverse beliefs have the smallest population. Finally, as observed in Tables 4 and 5, oil exporters are Islamic countries.
Table 6. Descriptive statistics of democracy and its determinants for 17 countries classified into six religious affiliations, 1996–2010
Notes: This table presents mean, median, maximum (max), minimum (min), and standard deviation (sd) values as well as number of observations (N) for different country groups. Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise.
Model
We examined the relationship between Islam and democratic adjustment by estimating the econometric model similar to that used by Barro (Reference Barro1999). We used control variables such as GDP, SCHOOL, GAP, URBAN, and POP, and introduced a dummy variable OIL. These control variables were included based on previous research. Democracy has been shown to be positively related to standard of living, which is directly proxied by GDP per capita, primary schooling, and income share of the middle class and inversely proxied by gender gap in primary schooling. In addition, democracy is negatively related to urbanization and dependency on natural resources (Barro, Reference Barro1999; Shafiq, Reference Shafiq2010). Information about oil-exporting countries was obtained from the Organization of Petroleum Exporting Countries.Footnote 8 Data for all of the other control variables are derived from the World Bank.Footnote 9
Considering that the sample covers 17 countries and spans the period from 1996 to 2010, a panel data model was used for estimation, and different specifications of the following econometric model were estimated.
$$\begin{equation*}
\begin{array}{l} \hspace*{-6pt}V{A_{i,t}}\,{=}\, {\beta _0}\,{+}\, {\beta _1}V{A_{i,t - 1}} + {\beta _2}V{A_{i,\ t - 1}} \times ISLAM {+} {\beta _3}\ln \left( {GD{P_{i,t - 1}}} \right) + {\beta _4}SCHOO{L_{i,t - 1}}\\ \quad \quad \ +\, {\beta _5}GA{P_{i,t - 1}} + {\beta _6}URBA{N_{i,t - 1}} + {\beta _7}\ln \left( {PO{P_{i,t - 1}}} \right) + {\beta _8}OI{L_{i,t}} + {u_i} + {v_{i,t}} \end{array}
\end{equation*}$$
where VA i, t is the democracy variable (voice and accountability) for country i in year t. VA i, t − 1 is the lagged dependent variable for country i in year t–1. The lag of the VA variable is included as one of the independent variables to estimate the speed of democratic adjustment, which is measured by (1 – the sum of the coefficient on VAi,t – 1 and that on VAi,t – 1 ×ISLAM) (Appendix A1). As a result, the larger (smaller) the coefficient on VAi,t – 1, the lower (higher) the adjustment speed. ISLAM is a dummy variable that returns a value of 1 if Islam is the major religion in a given country and 0 otherwise. GDP i, t − 1 is per capita GDP for country i in year t–1. SCHOOL i, t is primary schooling (number of years) for country i in year t–1. GAP i, t − 1 is the gap between male and female primary education (number of years) for country i in year t–1. URBAN i, t − 1 is the urbanization rate for country i in year t–1. POP i, t − 1 is the population for country i in year t–1. OIL i, t is the dummy variable that returns a value of 1 if country i is a member of OPEC in year t and 0 otherwise. ui is the country-specific effect for country i; v i, t is the white noise for country i in year t.
The fixed-effects panel data model was estimated to provide consistent estimators in hypothesis testing. Based on the preceding model, β 2 is expected to be negative if H2 holds up, that is, adjustment to the target level of democracy is faster in Islamic countries than in non-Islamic countries (Appendix A1).
4. Empirical results
Table 7 presents the results on the relationship between Islam and democracy. Panels A and B present the obtained results based on the entire sample and reduced sample, respectively. Focusing on Column 1 in Panel A, results on how democracy is related to its benchmark determinants generally concur with those in previous studies. The coefficients of GDP, URBAN, and OIL are insignificant, thus indicating that the level of democracy is generally unrelated to economic growth, urbanization, and being an oil exporter. The coefficients of the other three determinants are significant and demonstrate the expected signs. More specifically, SCHOOL has a significantly positive coefficient, indicating a positive relationship between the level of democracy and the duration of primary schooling, likely because education promotes civic participation. That is, people are more inclined to recognize the value of democracy and support democracy when they are more educated than when they are not (Glaeser et al., Reference Glaeser, Ponzetto and Shleifer2007). The coefficient of GAP is significantly negative, indicating a negative relationship of the level of democracy to the gender gap in primary schooling; that is, the level of democracy is higher when people of different genders are treated more equally in terms of education than when they are not. These results concur with the finding that gender equality is positively related to democracy (Inglehart et al., Reference Inglehart, Norris and Welzel2002). The coefficient of POP is significantly negative, indicating a negative relationship of the level of democracy to population. This observation supports the view that overpopulation can result in the dilution of democracy and transitive loss of democracy; that is, when a population exceeds a certain level, getting one's opinion heard by government becomes increasingly difficult as the population increases (Bartlett, Reference Bartlett2000). In sum, the results on the relationship between democracy and its benchmark determinants generally concur with those of previous research (Barro, Reference Barro1999).
Table 7. Relationship between Islam and democratic adjustment
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Focusing on the results on democracy, the coefficient on VAt –1 is significantly positive, thereby indicating that democratic adjustment is costly and that reaching the target level of democracy requires time. In addition, the coefficient on the interaction variable VAt –1×ISLAM is significantly negative, indicating that the speed of adjustment to the target level of democracy is higher for Islamic countries than for non-Islamic countries. These results support H2.
For the robustness check, we applied a stricter definition for Islamic countries in estimating the model to verify if the results in Column 1 continue to hold. Based on the religious demography provided by CIA World Factbook (CIA, 2011), Indonesia, Kuwait, and Malaysia have Muslim populations of less than 90% as opposed to other Islamic countries, such as Pakistan, Saudi Arabia, and Turkey, where close to or virtually 100% of the total population are Muslims.Footnote 10 The above-mentioned three countries were excluded incrementally from the Islamic country group and incorporated into the non-Islamic country group. When increasingly more countries were excluded from the Islamic country group in Columns 2 to 4, the results are similar to those in Column 1, except that the coefficient on POP becomes insignificant when Malaysia and Kuwait are excluded from the Islamic country group in Column 3. Most importantly, the coefficient on VAt –1×ISLAM remains significantly negative in Columns 2 to 4. Thus, the observation of a higher adjustment speed for Islamic countries in Column 1 remains robust when the Islamic country group is increasingly smaller in Columns 2 to 4.
Panel B presents the results obtained based on the reduced sample. Following the preceding approach, Islamic countries were increasingly excluded from the sample. However, different from Panel A, the excluded countries were not incorporated into the non-Islamic country group. The results on the relationship of democracy to its benchmark determinants are similar to those in Panel A, except that the coefficients on URBANt – 1 are borderline significant in Columns 1 and 2 of Panel B. Moreover, the coefficients on VAt – 1×ISLAM remain significantly negative in all of the columns of Panel B. Thus, the observation of faster democratic adjustment for Islamic countries from Panel A carries over to Panel B where the sample size decreases as increasingly more countries are excluded from the Islamic country group but not included in the non-Islamic country group.
To further confirm the robustness of the above-mentioned results, we divided the non-Islamic country group into five country groups with different belief types, and then reexamined the relationship of Islam to democracy by comparing Islamic countries to each of the other five non-Islamic country groups. Table 8 presents the results in four panels. Panels A, B, C, and D present the results obtained based on different combinations of Islamic countries; the panels consist of six, five, four, and three Islamic countries, respectively. In Panel A where the Islam country group consists of all six Islamic countries in the sample, the coefficient on VAt – 1×ISLAM is significantly negative in Column 3 where Islamic and Buddhist countries are compared, indicating that democratic adjustment is faster in Islamic countries than in Buddhist countries. In Panel B, where Malaysia is excluded from the Islamic country group, the coefficient on VAt – 1×ISLAM is significantly negative in Columns 2 and 3, where Islamic countries are compared to countries with diverse beliefs and Buddhist countries, respectively. The results indicate that democratic adjustment is faster in Islamic countries than in countries with diverse beliefs and in Buddhist countries. In Panel C, where Malaysia and Kuwait are excluded, and in Panel D, where Malaysia, Kuwait, and Indonesia are excluded from the Islamic country group, the results are similar to those in Panel A. That is, the coefficient on VAt – 1×ISLAM is significantly negative when Islamic countries are compared to the Buddhist country group in Column 3 of Panels C and D. The coefficient on VAt – 1×ISLAM is negative albeit insignificant when the Islamic country group is compared to Christian countries or the atheist country in Panels B to D of Table 8.
Table 8. Relationship between Islam and democratic adjustment – Islamic countries vs. countries with different belief types
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al. Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
In sum, the results strongly indicate that democratic adjustment is generally faster for Islamic countries than for non-Islamic countries. In addition, such a phenomenon is more pronounced when Islamic countries are compared to the Buddhist countries or countries with diverse beliefs.
Additional tests: interactions
To sustain democratic development for Islamic countries in the future, a higher target level of democracy should accompany faster democratic adjustment. Otherwise, faster democratic adjustment can imply faster reversion to any originally low level of democracy for Islamic countries. To this end, the study further estimated the model by interacting the Islam dummy variable (ISLAM) with each benchmark determinant of democracy at a time to provide policy recommendations on how to reach a higher target level of democracy more effectively for Islamic countries as opposed to non-Islamic countries. The corresponding coefficients are significant only when ISLAM interacts with URBANt – 1. Table 9 presents the results in seven columns, where Columns 1 to 4 present the results based on the entire sample, whereas Columns 5 to 7 present the results based on the reduced sample. The results in Table 9 indicate that the coefficient on the interaction variable URBANt – 1×ISLAM is significantly positive in all columns except Column 1 where the Islamic country group is the most comprehensive. Given that the coefficient on URBANt – 1 is insignificant in all columns, the relationship between urbanization and democracy is measured as the coefficient on URBANt – 1×ISLAM if a given country belongs to the Islamic country group and 0 otherwise. Therefore, results indicate a positive relationship between urbanization and democracy, suggesting that the target level of democracy is higher as the urbanization rate increases for Islamic countries only. Such results are robust to different coverages of Islamic countries and sample sizes; that is, the inclination of people to become increasingly receptive to different values and voices in urbanized society is verified for Islamic countries only.
Table 9. Relationship between Islam and democratic adjustment – interactions
Notes: Voice and accountability (VA) index is the proxy for democracy, ranging from –2.5 to 2.5 (Kaufmann et al., Reference Kaufmann, Kraay and Mastruzzi2011). Higher index values indicate higher level of democracy. ISLAM is the dummy variable that returns a value of 1 if a given country belongs to Islamic country group and 0 otherwise. GDP is per capita gross domestic product. SCHOOL is primary schooling (number of years). GAP is the gap between male and female primary education (number of years). URBAN is the urbanization rate. POP is population. OIL is the dummy variable that returns a value of 1 if a given country is a member of OPEC and 0 otherwise. In all columns, year dummies are included to capture year-specific effects, but the results are saved for brevity. N represents the number of firm-year observations; n stands for the number of countries. The numbers in the parentheses are cluster-robust standard errors. ***, **, *, and † indicates significance at .1, 1, 5, and 10% levels, respectively.
Based on the results in Table 9, the level of democracy is generally higher when education improves, gender gap narrows, population growth is under control, or countries are oil exporters. However, we demonstrate a unique relationship of democracy to urbanization for Islamic countries compared to non-Islamic countries. That is, results suggest that to reach a higher level of democracy, they might also want to focus on furthering urban development; this approach has been found to be ineffective for non-Islamic countries. Although our results differ from those of Barro (Reference Barro1999) that indicate no significant effect of urbanization on democracy, the study results are in line with recent findings that the level of democracy increases with urbanization, as well as modernization that is conducive to urbanization (Ciftci, Reference Ciftci2010; Inglehart and Welzel, Reference Inglehart and Welzel2010; Dima et al., Reference Dima, Leitão and Dima2011). Thus, our results likewise support the modernization theory in explaining democracy, although such theory gains support for Islamic countries only. That is, social-economic development plays a crucial role in promoting democracy effectively (Wucherpfennig and Deutsch, Reference Wucherpfennig and Deutsch2009).
Overall, the current study indicates that Islamic countries are generally less democratic than non-Islamic countries. However, the low level of democracy prevailing in Islamic countries likely results from not only Islam but also other Islam-related factors, such as culture and colonial experience. This study contributes to the existing literature by demonstrating that democratic adjustment is faster in Islamic countries than in non-Islamic countries, thereby suggesting that Islamic countries exhibit the potential to achieve considerable progress in democracy in the future. This study concludes that the relationship between Islam and democratic development is neither negative nor negligible, as documented in the majority of previous studies; rather, the study challenges the conventional wisdom that Islam impedes democratic development by providing strong empirical evidence indicative of faster democratic adjustment for Islamic countries. This finding provides hope for other Middle East countries that are recently undergoing democratic reforms but experiencing a faltering democratic transition. However, the observed faster democratic adjustment will be meaningful and constructive if Islamic countries simultaneously undertake actions to promote their target level of democracy. Otherwise, faster democratic adjustment can imply that Islamic countries can revert to any originally low level of democracy faster. The results further provide suggestions on how to effectively raise the target level of democracy for Islamic countries. Specifically, Islamic countries can release their democratic potential through urbanization to further promote democracy because results indicate a positive relationship between urbanization and the level of democracy for Islamic countries only, whereas no such relationship exists for non-Islamic countries.
5. Conclusion
Authoritarian governments have been successively overthrown and replaced by democratic regimes over the past few decades. In particular, Islamic countries that have been perennially regarded as undemocratic, such as those in the Arab world, recently launched a series of democratic revolutions. This move rekindled the democratic spirit and public awareness of democracy around the world. However, these Islamic countries may encounter difficulty in successfully developing democracy if their major belief, Islam, is truly undemocratic or anti-democratic in nature, as widely believed. Their democratic prospect is particularly challenging given that these Islamic countries are currently undergoing difficulties moving forward in democratization. Thus, this study primarily contributes to the existing literature by examining whether Islam hampers democratic development, as suggested by conventional wisdom. In contrast to previous research that is essentially retrospective and focuses on the static level of democracy in Islamic countries, this study looks into the future and adopts a dynamic approach by examining democratic adjustment speed in Islamic countries to provide a new insight into the perennially debated issue.
Using 17 Asian countries as the study sample, the results indicate that religion plays a role in determining democracy. Thus, future research on democracy should formally consider religion. Specifically, we demonstrate that the level of democracy in Islamic countries is lower than that in non-Islamic countries, consistent with the common belief and the findings of previous research. However, examining democracy from a dynamic perspective conveys an interesting story. Specifically, the level of democracy in Islamic countries exhibits an upward trend rather than a downward trend as observed for non-Islamic countries. Moreover, contrary to conventional wisdom that suggests the remarkable resilience of authoritarianism in Islamic countries, this study provides strong empirical evidence indicating that the democratic adjustment for Islamic countries is faster than that for non-Islamic countries, especially those with diverse beliefs and subscribing to Buddhism. Thus, the results refute conventional wisdom that Islam hinders democracy. Instead, Islamic countries are highly malleable and exhibit a potential for faster democratic development compared with non-Islamic countries. Given faster democratic adjustment, the level of democracy in Islamic countries should rapidly converge to that in other more democratic non-Islamic countries. In other words, if Islamic countries could accelerate their democratic momentum and continue on their democratic path, they would rapidly catch up with non-Islamic countries with higher levels of democracy. However, Islamic countries should upgrade their target level of democracy, such that they will be able to promote democracy to a new higher level more rapidly; otherwise, they might end up with the originally low level of democracy faster with the higher speed of democratic adjustment. Moreover, the study results provide policy implications on effectively increasing the target level of democracy. In addition to measures such as improving education, minimizing the gender gap, and controlling population growth, we demonstrate that Islamic countries likely own a unique advantage in increasing the target level of democracy by promoting urbanization. By contrast, urbanization appears to be irrelevant to democracy for non-Islamic countries.
Although a battery of robustness tests are conducted to ensure the validity of the study results, we recognize some limitations in this study and provide directions for future research. First, results presented in the study are based on 17 Asian countries. It is worthwhile to conduct further research using a similar approach to see if the study results carry over to other countries. Second, the study provides strong evidence indicating faster democratic adjustment for Islamic countries. However, although lagged variables are used as determinants of democracy in estimating the model to infer causal relations, the results are interpreted conservatively because endogeneity may not be well controlled. To make more reliable inferences about the causal relations on how to promote democracy effectively, more sophisticated econometric models can be employed in the future related studies. Third, the determinants of democracy used in the study are similar to those in prior studies. To shed new light on this research question, future research can explore whether other variables can also shape democracy for Islamic countries.
About the authors
Naiwei Chen is currently a professor of finance at Jiangxi Normal University, China. His research interests include political economy, corporate governance, international finance, Islamic banking and finance, corporate social responsibility, welfare, trade, and economic growth. He has published papers in Applied Economics, Applied Financial Economics, Asia-Pacific Financial Markets, Corporate Governance: An International Review, Economics of Governance, Journal of International Development, among others.
Tsai-Chen Yang was a masters student at National Chung Cheng University, Taiwan, under the guidance of Naiwei Chen.
Appendix A1. Panel data model specification
Assuming the target level of democracy (VA*) is a function of country-specific factors (X), the following equation is specified:
$$\begin{equation}
VA_{it}^* = \alpha + \mathop \sum \limits_j {\gamma _j}{X_{jit}} + {e_{it}}
\end{equation}$$
where i refers to country i; j refers to the jth country-specific factor; t refers to year t; eit refers to the error term for country i and year t. Further assume that countries are undergoing adjustment to reach their respective target level of democracy (VA*) based on the following partial adjustment mechanism:
$$\begin{equation}
V{A_{it}} - V{A_{i,t - 1}} = \rho \left( {VA_{it}^* - V{A_{i,t - 1}}} \right)
\end{equation}$$
where VAit is the actual VA whereas VA* it is the target VA, which can be estimated based on the above-mentioned econometric model (equation (1)). ρ is the adjustment coefficient that measures how fast VA adjusts to its target level, ranging from 0 to 1. If ρ = 0, VAit = VA i, t − 1, meaning that adjustment is costly and there is no way VA can revert to the target level. In contrast, if ρ = 1, VAit = VA* it , meaning that the adjustment incurs no cost and is instantaneous. The above equation can be rewritten as follows:
$$\begin{equation}
V{A_{it}} = \left( {1 - \rho } \right)V{A_{i,t - 1}} + \rho VA_{it}^*
\end{equation}$$
Substituting equation 1 into equation 3, we have:
$$\begin{equation}
V{A_{it}} = \delta V{A_{i,t - 1}} + \omega + \mathop \sum \limits_j {\theta _j}{X_{jit}} + {\varepsilon _{it}}
\end{equation}$$
where δ = 1 − ρ, ω = pα, θ j = ρrj and ε it = ρeit . Hence, the higher the value of δ, the lower the value of ρ and the slower the adjustment. In addition, testing H2 translates into testing whether the coefficient on VA t − 1 (δ) is smaller for Islamic countries than for non-Islamic countries.
